The present invention relates to systems for smoothing by spectral dispersion (SSD) of laser light, and particularly to systems for generating and spectrally dispersing broad-bandwidth light onto a phase plate having different elements which provide relative phase delays. The focused intensity pattern of the laser light then varies in time according to the frequency differences between the portions (beamlets) of the light incident on the elements to produce a time averaged intensity which will smooth.
This invention is especially suitable for use in providing laser beams which irradiate laser fusion targets wherein uniform absorption of laser light results in high-density compression, and avoids the generation of hot spot intensity non-uniformities. The invention will also be found useful wherever variations in the intensitY of a laser beam are not desired, for instance, when such variations can cause damage to the material through which the laser is propagating (and this damage is the result of the time integrated beam intensity).
This invention is an improvement over the invention of U.S. Pat. Application Ser. No. 228,131 filed Aug. 3, 1988 and assigned to the same assignee as this application. The prior application describes the general concepts of SSD. The subject matter of the prior application is also described in an article by R. S. Craxton, S. Skupsky and J. M. Soures, LLE Review, Volume 36, page 158 (1988).
The SSD concept is based upon the interference of laser light in a plane where, for example, a laser fusion target may be located. Consider the interference between rays from different elements of a phase plate which may be a distributed phase plate (DDP). In the target plane the combined electric fields from two exemplary rays are EQU E=E.sub.1 e.sup.i(kL.spsb.1.sup.+.phi..spsb.1.sup.-.omega.t) +E.sub.2 e.sup.i(kL.spsb.2.sup.+.phi..spsb.2.sup.-.omega.t), (1)
where the amplitudes are of the diffraction limited form E.sub.1 approximately equal to E.sub.2, approximately equal to sin(y)/y, and .phi..sub.1 and .phi..sub.2 are the phases of the rays (including those imposed by the phase plate). y is a transverse direction to the path of the light. For simplicity, if E.sub.1 equals E.sub.2, then the intensity variation is EQU I=.vertline.E.vertline..sup.2 =2E.sub.1.sup.2 +2E.sub.2.sup.2 cos[k(L.sub.1 -L.sub.2) +(.phi..sub.1 -.phi..sub.2)], (2)
which results in high-intensity fluctuations in the transverse (y) direction as the path length difference (L.sub.1 minus L.sub.2) from the phase plate to the target plane changes. If the rays have different frequencies and wave numbers (.omega..sub.1, k.sub.1) and (.omega..sub.2, k.sub.2), the intensity becomes EQU I=2E.sub.1.sup.2 +2E.sub.1.sup.2 cos]k.sub.1 L.sub.1 -k.sub.2 L.sub.2 +(.phi..sub.1 -.phi..sub.2) +(.omega..sub.1 -.omega..sub.2)t]. (3)
At any instant of time, the intensity pattern will still have high intensity modulations but they will fluctuate in time according to the frequency difference. When averaged over time, the interference term will approach 0 as 1/(.omega..sub.1 -.omega..sub.2)t, and the intensity will approach the smooth diffraction limited sinc.sup.2 envelope.
As described in the prior application and in the above-cited LLE Review article, it is necessary to assure that a plurality of frequencies are contained in the beam incident on the phase plate. Problems faced in practical implementation of an SSD system include (1) generation of bandwidth containing the plurality of frequencies that will not damage the laser glass or other optical elements in the system with high intensity spikes; (2) dispersion of the bandwidth frequencies across the DPP elements; (3) prevention of the distortion of the temporal profile of the beam; and (4) obtaining the smooth intensity pattern and improved uniformity during a sufficiently small averaging time to be useful in laser fusion and other applications.
Another constraint on SSD is to accommodate high-efficiency frequency tripling of the broadband light. Current laser fusion techniques use frequency tripled infrared (IR) light to take advantage of the increased collisional laser absorption and higher hydrodynamic implosion efficiency that is possible at shorter wavelengths. However, high efficiency tripling can only be achieved for a given orientation of the tripling crystals over a narrow spread in wave lengths (for example approximately one Angstrom in the IR). As described in the prior application and LLE review article, this can be accomplished by spectral dispersion of the beam in one direction and with cylindrical lenses to impose the required angular spread. It has, however, been difficult to obtain a linear spectral (wavelength) dispersion so as to obtain high efficiency frequency tripling.